{"id":"https://openalex.org/W6892012129","doi":"https://doi.org/10.48550/arxiv.2506.08966","title":"Pre-trained Language Models Learn Remarkably Accurate Representations of Numbers","display_name":"Pre-trained Language Models Learn Remarkably Accurate Representations of Numbers","publication_year":2025,"publication_date":"2025-06-10","ids":{"openalex":"https://openalex.org/W6892012129","doi":"https://doi.org/10.48550/arxiv.2506.08966"},"language":"en","primary_location":{"id":"doi:10.48550/arxiv.2506.08966","doi":"https://doi.org/10.48550/arxiv.2506.08966","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2506.08966","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"volume":null,"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2506.08966","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5032449871","display_name":"Marek Kadl\u010d\u00edk","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Kadl\u010d\u00edk, Marek","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102979706","display_name":"Michal \u0160tef\u00e1nik","orcid":"https://orcid.org/0000-0003-1766-5538","observed_orcids":["https://orcid.org/0000-0003-1766-5538"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"\u0160tef\u00e1nik, Michal","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5073653815","display_name":"Timothee Mickus","orcid":"https://orcid.org/0000-0001-9538-7209","observed_orcids":["https://orcid.org/0000-0001-9538-7209"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Mickus, Timothee","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5104202378","display_name":"Michal Spiegel","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Spiegel, Michal","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5139804152","display_name":"Josef Kucha\u0159","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Kucha\u0159, Josef","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":0,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":null,"last_page":null},"is_retracted":false,"is_paratext":false,"is_xpac":true,"primary_topic":{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.9179999828338623,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.9179999828338623,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10181","display_name":"Natural Language Processing Techniques","score":0.017999999225139618,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T12026","display_name":"Explainable Artificial Intelligence (XAI)","score":0.013500000350177288,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/pre-trained-language-models","display_name":"pre-trained language models","score":0.8840000033378601},{"id":"https://openalex.org/keywords/calculation-errors","display_name":"calculation errors","score":0.7749999761581421},{"id":"https://openalex.org/keywords/probing-technique","display_name":"probing technique","score":0.5590000152587891},{"id":"https://openalex.org/keywords/sinusoidal-pattern","display_name":"sinusoidal pattern","score":0.7789999842643738},{"id":"https://openalex.org/keywords/number-representation","display_name":"number representation","score":0.375}],"concepts":[{"id":"https://openalex.org/C2778858076","wikidata":"https://www.wikidata.org/wiki/Q5249539","display_name":"Decodes","level":3,"score":0.7046999931335449},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6211000084877014},{"id":"https://openalex.org/C204323151","wikidata":"https://www.wikidata.org/wiki/Q905424","display_name":"Range (aeronautics)","level":2,"score":0.6054999828338623},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.47780001163482666},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.4316999912261963},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.40959998965263367},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.37049999833106995},{"id":"https://openalex.org/C2983448237","wikidata":"https://www.wikidata.org/wiki/Q1078276","display_name":"Language understanding","level":2,"score":0.36629998683929443},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.36399999260902405},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.31310001015663147},{"id":"https://openalex.org/C18762648","wikidata":"https://www.wikidata.org/wiki/Q42213","display_name":"Work (physics)","level":2,"score":0.3084999918937683},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.2935999929904938},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.27790001034736633},{"id":"https://openalex.org/C94375191","wikidata":"https://www.wikidata.org/wiki/Q11205","display_name":"Arithmetic","level":1,"score":0.263700008392334},{"id":"https://openalex.org/C2780586882","wikidata":"https://www.wikidata.org/wiki/Q7520643","display_name":"Simple (philosophy)","level":2,"score":0.25920000672340393},{"id":"https://openalex.org/C116834253","wikidata":"https://www.wikidata.org/wiki/Q2039217","display_name":"Identification (biology)","level":2,"score":0.2574000060558319}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2506.08966","doi":"https://doi.org/10.48550/arxiv.2506.08966","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2506.08966","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2506.08966","doi":"https://doi.org/10.48550/arxiv.2506.08966","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2506.08966","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"sustainable_development_goals":[],"sustainable_development_goals_aurora":[{"id":"https://metadata.un.org/sdg/4","display_name":"Quality education","score":0.7863473892211914}],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Pretrained":[0],"language":[1],"models":[2],"(LMs)":[3],"are":[4,47],"prone":[5],"to":[6,28],"arithmetic":[7],"errors.":[8,134],"Existing":[9],"work":[10],"showed":[11],"limited":[12],"success":[13],"in":[14,36,116],"probing":[15,45,66],"numeric":[16,70],"values":[17,71],"from":[18,72],"models'":[19],"representations,":[20],"indicating":[21],"that":[22,43,68,86,100,121],"these":[23,133],"errors":[24,115],"can":[25,131],"be":[26],"attributed":[27],"the":[29,50,88,101,123,126],"inherent":[30],"unreliability":[31],"of":[32,53,81,113],"distributionally":[33],"learned":[34,54],"embeddings":[35,56,74,124],"representing":[37],"exact":[38],"quantities.":[39],"However,":[40],"we":[41,62,98],"observe":[42],"previous":[44],"methods":[46],"inadequate":[48],"for":[49],"emergent":[51],"structure":[52],"number":[55],"with":[57,75,94,125],"sinusoidal":[58],"patterns.":[59],"In":[60],"response,":[61],"propose":[63],"a":[64,79,110],"novel":[65],"technique":[67],"decodes":[69],"input":[73],"near-perfect":[76],"accuracy":[77],"across":[78],"range":[80],"open-source":[82],"LMs.":[83],"This":[84],"proves":[85],"after":[87],"sole":[89],"pre-training,":[90],"LMs":[91],"represent":[92],"numbers":[93],"remarkable":[95],"precision.":[96],"Finally,":[97],"find":[99],"embeddings'":[102],"precision,":[103],"judged":[104],"by":[105],"our":[106,128],"probe's":[107],"accuracy,":[108],"explains":[109],"large":[111],"portion":[112],"LM's":[114],"elementary":[117],"arithmetic,":[118],"and":[119],"show":[120],"aligning":[122],"pattern":[127],"probes":[129],"discover":[130],"mitigate":[132]},"counts_by_year":[],"updated_date":"2026-10-14T16:39:01.820549","created_date":"2025-10-10T00:00:00"}